Diagnosis constitutes a major categorisation tool in medicine. This paper adds to the paucity of knowledge about part of the work such a tool performs. It examines the ways in which diabetes epidemiology translates into attempts to prevent diabetes from occurring through screening and diagnosing, and then managing those categorised as ‘at high risk’. Using a qualitative design in the context of a small-scale Danish intervention study, the findings suggest that health professionals overinterpreted relative risk reductions and exaggerated treatment effects; simultaneously, prediabetics called into question the scientifically set thresholds. Nonetheless, arguments concerning changed behaviour’s benefits were sought to be incorporated into self-care routines. The data highlight how, by identifying a normatively set threshold, a diagnosis can determine health promotion messages’ impact in ‘high-risk’ prevention strategies.